1911-12.] Transverse Induction Changes in Demagnetised Iron. 311 
In both figs. 11 and 12 the curves on the left ( — H c ) cross the 
original curves of complete demagnetisation and each other as the residual 
magnetisation is increased. In fig. 10 this crossing is sufficiently well 
marked, and is in fact an early stage of the large initial rise (figs. 11 and 12) 
from the third to the second quadrant which occurs when the residual 
magnetisation is sufficiently increased and the subsequent magnetising 
field negative. 
On the other hand, the curves on the right do not appear to cross, 
fig. 9 for weak magnetisation being in this respect typical of the whole 
series. Attention, however, may be called to the fact that when the 
residual magnetisation is small, the curves nearly coincide with that 
of complete demagnetisation which may indicate crossings close to 
the origin. 
When 0 = 90° (the curves of complete demagnetisation having vanished 
in the horizontal axis) small initial crossings are obtained, and to show 
them the ordinates have been expanded one thousand times in fig. 13. Note 
that the curves on the right and left are now symmetrical about the 
vertical axis. 
The residual magnetisation for each pair of curves ( + H and — II ) 
was not directly measured (figs. 11, 12, and 13), hut may he approximately 
obtained from the consideration that the final negative values of the 
various curves, when H c is sufficiently increased, is equal to the residual 
magnetisation into sin 0 — the angle at which partial demagnetisation 
had been performed. 
It may be mentioned that when the steel tube is left residually 
magnetised in the ordinary way , the crossings of the curves when H c is 
negative (figs. 11 and 12) are less marked and their rise into the second 
quadrant either entirely absent (0 = 7l o, 6) or very small (d = 18 0, 6) — an 
entirely similar result to that obtained in the previous investigation. 
The simple withdrawal of the positive field leaves the iron more sus- 
ceptible to negative than positive change. On the other hand, if the steel 
tube be left residually magnetised by the withdrawal of the positive 
field, followed by the application and withdrawal of a reduced negative 
field, the iron may be left positively magnetised but more susceptible to 
positive than to negative change. Under this condition it was found 
that the rise of the transverse curves into the second quadrants is readily 
obtained. The production of the residual magnetisation by the method 
of partial demagnetisation adopted in the present investigation eliminates 
both these experimental extremes — the result of the more complex form 
of magnetic seolotropy already referred to. 
